AI Art Weekly #118

Hello, my fellow dreamers, and welcome to issue #118 of AI Art Weekly! 👋

Google released a new experimental multimodal model this week called Gemini 2.0 Flash.

Now multimodality doesn’t sound like anything special these days, but so far we never really had the ability to experience true multimodal output ourselves. Models could process images, but they couldn’t produce them directly without calling another model in the background. Gemini 2.0 can directly generate images, and the consequences are quite wild.

With the ability to directly understand and generate images, the model can do any image editing task imaginable: inpainting, outpainting, style transfer, identity transfer, pose transfer, even generate multiple frames for gifs – the list goes on. I compiled a few examples in an X thread this week. I strongly feel that models like this, when scaled up, are going to be the real game changers. Just imagine seamless conversion from any modality to another: text, images, 3D, videos, audio, or even brainwaves. The possibilities will be endless.

Even though things have been pretty wild these last 3 years, we’re still so early, barely scratching the surface with LLMs. Strap in, because the AI train isn’t going to stop anytime soon.

PROMPTCACHE

Support the newsletter and unlock the full potential of AI-generated art with my curated collection of 275+ high-quality Midjourney SREF codes and 2000+ creative prompts.

promptcache.com